HEAT TREATMENT. 309
this does not always represent the temperature exactly unless the "invisible" condition is obtained.
No. 16 represents the micro-structure of a steel casting unan-nealed, magnified 20 diameters. It is almost impossible to give an idea of the structure in a small photograph, but the illustration shows parts of three grains, and like all the other reproductions, is typical. No. 17 shows the same casting after annealing. The picture is not all it should be, but by careful examination a remarkably small grain may be distinguished; the areas of pearlite and ferrite are indicative of an insufficient breaking up of the microscopic forms. No. 18 represents the casting after a second annealing. No, 14 and No. 15 show the structure before and after annealing of a special high carbon casting used in railroad work where ability to withstand shock is of prime importance. .A As stated in Section XVi, the second critical point is character-" ized by a loss of the magnetic properties in heating; this point is very easily determined by using an electro magnet, the wires of which are connected with a sensitive galvanometer. The act of moving the magnet into and away from contact with the metal moves the needle of the galvanometer 'as long as the metal is magnetic. It would seem as if this should be a good point to agree upon as the temperature to which castings shall be heated for annealing. Sufficient data are not available to state positively that such treatment would give the best results possible, but it seems quite certain that treatment on this line would give good structure and be a great improvement on most of the haphazard methods now in use.
SEC. XVn.—Effect of heat treatment on the structure of rolled material.—In order to determine the effect of heat treatment on the structure of rolled material, tests were taken from finished angles, the general method of procedure being as follows:
A piece five feet long was sheared from the angle and cut into five equal lengths. An ordinary test bar was taken from one of the legs of each piece in the same relative place and numbered from 1 to 5. From each of the extremes 1 and 5 a section was cut for the microscope and the bars pulled in the testing machine to prove that the piece was homogeneous. The bars, 2, 3 and 4, were treated in a muffle heated by an electric coil at temperatures varying from 625° C. to 890° C., the temperature in all experiments being taken by a Le Chatelier pyrometer. No attempt was made to heat